New light on the formation and evolution of bars: Trends in the stellar line-strength indices distribution inside the bar region

0.15 MB·English

by  I. Perez

Astronomy&Astrophysicsmanuscriptno.abundance˙perex (cid:13)c ESO2008
February5,2008
LE
New light on the formation and evolution of bars
⋆
Trends in the stellar line-strength indices distribution inside the bar region
I.Pe´rez1⋆⋆,2⋆⋆⋆,P.Sa´nchez-Bla´zquez3†,andA.Zurita2‡
1 KapteynAstronomicalInstitute,UniversityofGroningen,theNetherlands
7 email:[email protected]
0 2 DepartamentodeF´ısicateo´ricaydelCosmos,UniversidaddeGranada,Spain
0 email:[email protected]
2 3 CentreForAstrophysics,UniversityofCentralLancashire,UK
email:[email protected]
n
a
ReceivedSeptember15,1996;acceptedMarch16,1997
J

9
ABSTRACT
1

Aims.Theaimistostudythestellarcontentofthebarregiontoconstrainitsformationandevolution.
2
Methods.Line-strength indices in the bar region of a sample of 6 barred galaxies were employed to derive age and metallicity
v gradientsalongthebarsusingstellarpopulationmodels.
9 Results.Wefindclearradialgradientsintheline-strengthindicesforallthegalaxies.Wefindpositivegradientswithinthebarregion
5 inthemetalindicesinfourofthesixgalaxies,andoppositetrendsintheothertwo.TheselattertwogalaxiesareclassifiedasSAB
1 andtheypresentexponentialbarlightprofiles.ForallthegalaxieswefindapositivegradientintheBalmerindices.Thereisaclear
2 correlationbetweenthepositionofmorphologicalfeaturesinsidethebarregionwithchangesintheslopeandvalueoftheindices,
1 whichindicate,usingstellarpopulationanalysis,changesinthestellarpopulations.Therefore,itseemsthatthebarregionsshowa
6 gradientinbothageandmetallicity,changingradiallytoyoungerandmoremetalrichpopulationsforallthegalaxiesexceptforthose
0 twowithexponentialprofiles.
/
h
Keywords.Galaxies:abundances–Galaxies:formation–Galaxies:evolution
p
-
o
r 1. Introduction ing (Binney&Tremaine, 1987), so one would not expect the
t
s sameabundancetrendsinbothcomponents.Friedli(1998)pre-
a The importance of bars in galaxies as mechanisms to transfer
dicted, using N-body simulations of bars with pre-existing ex-
: angularmomentumandmatteronlargescaleswithinthegalaxy
v ponentialabundances,a nullevolutionofthe stellar abundance
is indisputable (Kormendy&Kennicutt, 2004, and references
i profile(althoughwithadecreaseinthemeanmetallicity)while
X therein).Barsarethoughttobeformedthroughspontaneousdisk
thegasabundanceprofileflattenedrapidly.Asithasbeenmen-
r instabilities or by instabilities produced during galaxy encoun- tionedbefore,thegas-phaseabundancedistributionpredictedby
a
ters;thedetailsofbarformationandtheirsubsequentevolution
Friedli (1998) has been shown to agree with the observational
remains a matter of debate. Until now, only detailed analyses
data of abundances in HII regions. As for the stellar compo-
of the gas-phase abundancesalong bars have been undertaken,
nent, there has not been any study addressing the stellar pop-
from observations of emission lines of HII regions. Studies of ulation along bars using spectroscopic information. Recently,
the gas-phase abundances provide us with a present-day snap-
Moorthy&Holtzman(2006)havepresentedresultsforthestel-
shot of the interstellar medium abundance. These studies have
lar populationsof bulges of spiral galaxies, some of which are
shown that there is little variation in the chemical abundances
barred galaxies, finding a difference between bulges of barred
in the gas-phase along the bar (Martin&Friedli, 1997, 1999).
andun-barredgalaxies.
Thestudyofthestellaragesandmetallicities,ontheotherhand,
givesus’archaeological’cluesastotheformationandevolution Todate,ithasbeendifficulttoobtainradialabundancedis-
of the bar. Furthermore, the evolution of the two components, tributionsalong barsbecause, in orderto obtain reliable stellar
gasandstars,suffersfromverydifferentevolutionaryprocesses; abundancesandages,veryhighsignal-to-noise(S/N)spectraare
the gas is mainly dominated by the gravitational torque of the necessary, and althoughbars are high surface brightnessstruc-
non-axisymmetricmass component, while the evolution of the tures, this still implies long integration times on medium-size
stellar component is mainly affected by different orbital mix- telescopes.WehavecarriedoutaprojecttoobtainhighS/Ndata
alongthebarmajoraxisofasampleofbarredgalaxies.Thisis
⋆ Based on observations obtained at Siding Spring Observatory the first detailed study of stellar line indices along the bar re-
(RSAA,ANU,Australia) gion.Inthisletter,we showclearevidenceforline-strengthin-
⋆⋆ VeniFellow dexgradientsalongthebarsandacorrelationbetweentheposi-
⋆⋆⋆ AssociateResearcher tionofmorphologicalfeaturesinsidethebarregion(bulge/inner-
† MarieCurie disk/inner-bar/primary-bar)andbreaksintheirrespectivestellar
‡ RetornoJ.AFellow populations.
2 I.Pe´rezetal.:Newlightontheformationandevolutionofbars
2. Observationsanddatareduction measured and are presented in Fig. 1. We adopted the defi-
nition of Trageretal. (1998) and Worthey&Ottaviani (1997)
2.1.Sampleselection
for the high-order Balmer lines Hδ and Hγ. The Fe3n index
We have selected barred galaxies from the Third Reference derived here is a combination of three prominent Fe lines (
Catalogueofbrightgalaxies(RC3)(deVaucouleurs,1948),with Fe4383+Fe5270+Fe5335)/3), similar to the Fe3 indexdefined
the following criteria; to be southern (δ < 0◦), classified as byKuntschner(2000),withFe4383insteadofFe5015.
barred and with inclinations between 10◦ and 50◦, and nearby To compare between different regions of the galaxy and to
(cz≤ 2500kms−1) to be able to properlyresolve the bar,with compare with other galaxies, indices have to be measured at
morphological types earlier than SBb to avoid morphological the same resolution. As to allow flexibility in the model selec-
dependencyof the results (in this first approach)and to ensure tionandtopotentiallycomparewiththemeasurementsofother
thattherequestedS/Nisreached,sinceearly-typegalaxieshave galaxycomponentsfromotherstudies,wedecidedtofollowthis
higherbarsurfacebrightness,crucialfortheline-strengthdeter- approach.AfterbroadeningtotheLickresolution,line-strength
mination.Wefinallyobservedsixgalaxies(namesandmorpho- indicesweremeasuredusingtheINDEXroutine(Cardiel,1999).
logicaltypesareindicatedinFig.1).Itisbynomeansastatis- To derive errors in the line-strength indices, we followed a
ticallycompletesample,butitisastartingpointinthestudyof similar approach as in Kelsonetal. (2006). This method con-
line-strengthindicesinbars. sistsincalculatingtheerrorsinthemeanfluxwithinapassband
using noise spectra derived by subtracting a best-fit model to
thegalaxyspectra.Toderivethebest-fitweusedtheOPTEMA
2.2.Observationsanddatareduction algorithmdescribedby Gonza´lez(1993).We used the synthetic
spectralenergydistributionsprovidedbyVazdekisetal.(2007).
We obtained long–slit spectra of six barred galaxies with the
Oncetheerrorinthe meanfluxofeachbandpassiscalculated,
Double Beam Spectrograph (DBS) on the 2.3m telescope at
the formulae derived in Cardieletal. (1998) were used to cal-
Siding Spring Observatory (SSO) during February 2006. SITe
culatetheerrorsintheindices.Asthevaluesoftheindicesde-
1752×532 CCD detectors were used. The gratings employed
pendsontheDopplerbroadening,thevelocitydispersionalong
werethe600Bandthe600Rfortheblueandredarms,respec-
the radii was also calculated. To derive σ and radial veloci-
tively,withaslit-widthof2arcsec.Thisset-upgivesadispersion
ties, we use the MOVEL and OPTEMA algorithms described
of1.1Å/pixelforthebluearmand1.09Å/pixelfortheredarmin
byGonza´lez(1993).To buildan optimaltemplatewe use stars
thewavelengthintervals3892–5814Åand5390–7314Årespec-
observedwithexactlythesameinstrumentalconfigurationasthe
tively,givingavelocitydispersionresolution∆σ ≈ 130kms−1.
galaxies.Tocorrecttheindicesforthevelocitydispersionbroad-
The slit was placed along the bar major axis. The slit length is
eningwefollowedthestandardprocedureofderivingcorrection
7 arcmin, which allows one to study nearby galaxies in detail,
curves from artificially broadened spectra. We derive, for each
while still providing sufficient sky coverage. The position an-
galaxy, a different curve using the best-fit spectra obtained in
gle was derived using the DSS images of the sample galaxies.
the derivation of the errors. The reason to derive a polynomial
Comparison arc lamp exposures were obtained for wavelength
foreachgalaxyisthatthebroadeningcorrectiondependsonthe
calibration.Spectrophotometricstandardswereobservedwitha
strength of the indices (Kuntschner, 1998), and by adopting a
slit width of 6 arcsec. Additionally, we observed 11 G-K stars
single polynomialfor all the galaxies, artificial trends between
from the Lick/IDS library to be used as templates for velocity
indicesandσcanbeintroduced(Kelsonetal.,2006).
dispersion (σ) measurements, as well as to transformour line-
Tocorrecttheindicesforanynebularemissioncontribution
strength indices to the Lick system. The total integration time
weusedGANDALF(Sarzietal.,2006)whichisasimultaneous
foreachofthegalaxieswastypically3hours.
emissionandabsorptionlinesfittingalgorithm.Wehavemarked
Thedatapresentedinthisletterreferstothebluerangealone.
all the points with detected emission (see Fig. 1), where the
All the spectra were reduced using standard IRAF1 routines.
detection threshold has been calculated with the prescriptions
Overscanandbiasweresubtractedfromalltheframes.Nodark
given in Sarzietal. (2006). The sensitivity vary from 0.3 Å in
subtraction was done due to the low dark current of the chips
the internal parts to 0.6 Å in the external parts of the galaxy.
used.Flatfieldingcorrectionwasachievedto the2%level.The
When the emission completely fills all the absorption Balmer
relatively small, compared to the slit length, size of the galax-
lines,thecorrectionismoreuncertain.Forthosecases,wehave
iesallowedforaccurateskysubtractionforeachframepriorto
useddifferentsymbolsinFig.1.
combiningthereducedspectra.Afirstorderpolynomialwasfit
WeusedTMB03,Bruzual&Charlot(2003)andVazdekiset
alongthespatialdirectionandthensubtractedfromtheframes.
al. (2007) models to derive simple stellar population ages and
Spectrophotometricstandards were used to relatively flux cali-
metallicities using the indicesHβ and [MgbFe]′ (TMB03). We
brate the spectra. For each fully reduced galaxy frame, a final
also derived [α/Fe] using Fe4383 and Mgb by fixing the ages
frame was created by extracting spectra along the slit, binning
obtainedinthe[MgbF]′vs.HβdiagramandtheTMB03models.
in the spatial directionto guaranteea minimumsignal-to-noise
ratioof20perÅinthespectralregionofMgb.
4. Locationofthemorphologicalfeatures
3. Line-strengthindices
Inordertolocatetheradiusofchangesinmorphologicalstruc-
ture,ellipsefittingtothelightdistributionwasperformedusing
In order to analyse the stellar population of the bars, spec-
tral indices of the Lick/IDS system (Faberetal., 1985) were theIRAFtaskELLIPSE.Allthegalaxiesshowedpoint-likenu-
clear regions,so thecentrewas fixedusingthe coordinatesob-
1 IRAF is distributed by the National Optical Astronomy tained by fitting a Gaussian to the nucleus. The position angle
Observatories,whichisoperatedbytheAssociationofUniversitiesfor (PA) and the ellipticity (e) were left as free parameters in the
Research in Astronomy, Inc. (AURA) under cooperative agreement fitting. Changes in both ellipticity and PA were considered as
withtheNationalScienceFoundation. candidate structures, then visual inspection of the images was
I.Pe´rezetal.:Newlightontheformationandevolutionofbars 3
performedtoseeiftherewasanyanomaly.Theanalysisissim- There is an obvious influence of the different components
ilar to thatdescribedin Erwin (2004). The imagesused forthe tothe shapeoftheradialdistribution.We havemadetestssub-
analysiscomefromdifferentsources,allthe analysiswasdone tractingabulgecomponent(fromthedecompositionofthelight
with 2MASS data and also a different band in the cases were profile),thisonlyhelpstoobtainanevenclearerresultforthebar
there werepublishedimages. Thefollowingsourceswere used andinnerstructure,withoutchangingthebasictrends.Weprefer
for each of the galaxies: for NGC 3081 the HST image of the toshowthecompleteradialprofileinordertoavoidaffectingthe
nuclear region in the F606W band (Malkanetal., 1998), for theobservedprofileswithanymodel-dependentanalysis.
NGC1433the R-bandbyHameed&Devereux(1999),andfor Independently of the selected models (see Section 3), we
NGC 5101 B-band image by Eskridgeetal. (1999). We com- derive positive age and metallicity gradients in the bars (outer
paredthe resultsin the differentbandsfinding goodagreement partsofthebarsbeingyoungerandmoremetalrichthanthein-
inthepositionofthemorphologicalfeatures. ternal parts), exceptfor the two galaxieswith exponentialbars
Forthreeoutofthesixgalaxies(NGC1433,NGC3081,and (NGC2665andNGC 3081).In a futurepaperwe plan to do a
NGC2217)therearepublisheddataontheradiiofmorpholog- moredetailedanalysisofthestarformationhistoryandchemical
ical structures within the inner region (Erwin, 2004). There is evolutionofalltheobjects.There,wewillpresentthecompari-
goodagreement(differencesof≈5%)betweenouranalysisand sonoftheseresultswithdynamicalandstellarpopulationmod-
thevaluesfoundbyErwin(2004). els using the entire spectra to try to interpret the results in the
contextofbarformationandevolution.
Summarising, we have obtained high S/N long-slit spectra
5. Resultsandsummary
alongthebarsofsixearly-typebarredgalaxiesandwehavede-
Figure 1 shows the radial trendsof two indices for each of the rived radial profiles of the line-strength indices for the sample
galaxies, a metal and a Balmer index. The other Balmer in- galaxies. There is a clear difference between the indices in the
dicesfollowthesametrendsshowninFig.1.TheMgindicesin barregion(inFig.1, theregionbetweenthe twoverticallines)
somecasesdifferfromtheFeindices,followingoppositetrends. andthecentralcomponent.Withinthebarregionthereisaclear
However, one has to bear in mind that raw line indices are af- gradient in both metal and Balmer lines that when combined
fected by the age–metallicity degeneracy. Therefore, to disen- with stellar population models for all galaxies gives a gradient
tangleageandmetallicityeffectsfromtheindices,onerequires inboth,metallicityandage,indicatingthatouterpartsofthebar
the use of stellar population models. In the cases where Hβ is areyoungerandmoremetalrich.Theseresultsmightindicatea
moreaffectedbyemissionHδwaschosen,sinceitislessaffected secularinside-outgrowthforthebar.
bytheemissionlines(seeSection3aboutemissioncorrection).
Acknowledgements. WearegreatfultoReynierPeletierforallowingustocom-
OneshouldbecarefulderivingtheBalmerindicesaffectedwith
parewithhiscodetheemissioncorrection.WearereallygreatfultoB.Gibson
emission since the derived values are sensitive to the emission andK.Gandaforthecarefulreadingofthemanuscript.Wethankthereferee,
correction (increasing the value after the correction). We have A.Vazdekis,fortheveryusefuldiscussionsandcommentsonthemanuscript.
triedseveralmethods(evennocorrection)andthetrendsdonot This publication makes use of data products from the Two Micron All Sky
change.InFig.1onecanclearlyseeadifferenttrendofthera- Survey, which is a joint project of the University of Massachusetts and the
Infrared Processing and Analysis Center/California Institute of Technology,
dialdistributionoftheindiceswithachangeattheradiusofthe
fundedbytheNationalAeronauticsandSpaceAdministrationandtheNational
innerstructure(inFig.1,themostinnerverticalline). Science.I.P.acknowledgesfinancialsupportfromtheNetherlandsOrganisation
Forclarity,wedividetheresultsoftheradialdistributionin: forScientific Research (NWO)Foundation andtheLeidsKerkhoven-Bosscha
a)theinnerstructure,uptothemostinnerverticallineinFig.1, Fonds.A.Z.acknowledgessupportfromtheConsejer´ıadeEducacio´nyCiencia
delaJuntadeAndaluc´ıa,Spain.
andb)thebarregion,betweenthetwoverticallinesinFig.1.
a) Inner region. There seems to be a gradient of the metal
andBalmerlineswithintheinnerstructure,withouttakinginto References
accountthecentralpoints.Theseverycentralpointsarepossibly
dominatedbythebulge,andshowadifferentbehaviourfromthe Binney,J.&Tremaine,S.1987,GalacticDynamics(Princeton:PrincetonUniv.
Press)
bar and the inner structure. Both sides of the bar show similar
Bruzual,G.&Charlot,S.2003,MNRAS,344,1000
trends,reinforcingtheresults. Cardiel,N.1999,Ph.D.Thesis
b) Bar region. Within the bar region the slope changes for Cardiel,N.,Gorgas,J.,Cenarro,J.,&Gonza´lez,J.1998,A&AS,127,597
both the metal and the Balmer indices. For NGC 5101 and Combes,F.&Elmegreen,B.1993,A&A,271,391
deVaucouleurs,G.1948,Ann.Astrophys.,11,24
NGC 4643, there seems to be a point where the slope changes
Erwin,P.2004,MNRAS,364,283
againcorrespondingtotheradiusaroundthepointofmaximum Eskridge,P.,Frogel,J.,Pogge,R.,etal.1999,AJ,118,730
ellipticity, which is not considered as the end of the bar in the Faber,S.M.,Friel,E.D.,Burstein,D.,&Gaskell,C.M.1985,ApJS,57,711
ellipticity-PA analysis. NGC 1433 shows an Fe3n distribution Friedli,D.1998,in,287
Gonza´lez,J.J.1993,Ph.D.Thesis
also compatiblewith no radialchangein the Fe3n, but there is
Hameed,S.&Devereux,N.1999,AJ,118,730
acleardifferencebetweentheinnerpartandthebarregion.The
Kelson,D.D.,Illingworth,G.D.,Franx,M.,&vanDokkum,P.G.2006,ArXiv
Balmerindicesshowapositivegradientandachangecompared Astrophysicse-prints
to the innerregion.For two cases (NGC 2665and NCG 3081) Kormendy,J.&Kennicutt,R.2004,ARA&A,42,603
there seems to be a negative trend of the metal indices and a Kuntschner,H.1998,Ph.D.Thesis
Kuntschner,H.2000,MNRAS,315,184
slightlynegativetrendintheBalmerindices;theydonotshowa
Malkan,M.,Gorjian,V.,&Tam,R.1998,ApJS,117,25
changeintheslopeinthebarregion;interestinglyenoughthese Martin,P.&Friedli,D.1997,A&A,326,449
are the only two galaxiesclassified as SAB in the sample, and Martin,P.&Friedli,D.1999,A&A,346,769
theypresentanexponentialprofileinthebar,whichmightindi- Moorthy,B.&Holtzman,J.2006,MNRAS,371,583
cateadifferentformationprocessandhistorycomparedtothose Noguchi,M.1996,ApJ,469,605
Sarzi,M.,Falco´n-Barroso,J.,Davies,L.,etal.2006,MNRAS,366,1151
withaflatprofile(Combes&Elmegreen,1993;Noguchi,1996).
Trager,S.C.,Worthey,G.,Faber,S.M.,Burstein,D.,&Gonzalez,J.J.1998,
Thispointdeservesfurtherinvestigation.Forallgalaxies,there ApJS,116,1
isalsogoodagreementbetweenbothsidesofthebar. Worthey,G.&Ottaviani,D.L.1997,ApJS,111,377
4 I.Pe´rezetal.:Newlightontheformationandevolutionofbars
Fig.1.Radialdistributionoftwoindicesforeachofthegalaxies(acombinedFeindexandtheH indexortheH imagedepending
β δ
on the emission in the bin, see text). The open circles and the filled triangles represent both sides of the bar. We have marked
all the points with detected emission with a star, where the detection threshold has been calculated with the prescriptions given
in Sarzietal. (2006). When the emission completely fills all the absorption Balmer lines, the correction is more uncertain. For
thosescases,wehavemarkedthepointswithanarrow.Theverticaldashedlinesrepresenttheend-pointofmorphologicalfeatures,
meaning the inner bar or nuclear disk and the primary bar. Notice the general change in the indices for the two regionsand the
gradientsinboth.